record-over-test_20260525_183326

该数据集是一个用于机器人任务的数据集，由LeRobot创建。它包含机械臂控制相关的动作和观察数据，具体特征包括：动作（action）和观察状态（observation.state），均涉及六个关节位置（肩部平移、肩部提升、肘部弯曲、手腕弯曲、手腕旋转和夹持器位置）；观察图像（observation.images.front）为前视摄像头视频，分辨率为480x640，3通道，帧率为30fps；其他特征包括时间戳（timestamp）、帧索引（frame_index）、剧集索引（episode_index）、索引（index）和任务索引（task_index）。数据集结构基于parquet文件，总剧集数为1，总帧数为299，总任务数为1，数据文件大小为100MB，视频文件大小为200MB。机器人类型为so_follower，数据分割为训练集（从0到1）。该数据集适用于机器人学习和控制任务，特别是机械臂操作和视觉感知。This is a robotics task dataset created by LeRobot. It encompasses action and observation data related to robotic arm control, with specific features as follows: Action and observation.state both involve six joint positions: shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation and gripper position; observation.images.front is the front-facing camera video with a resolution of 480×640, 3 channels and a frame rate of 30 fps; Additional features include timestamp, frame_index, episode_index, index and task_index. The dataset is structured based on Parquet files. The total number of episodes is 1, total frames are 299, total tasks are 1, the data file size is 100 MB and the video file size is 200 MB. The robot type is so_follower, and the data split is the training set ranging from 0 to 1. This dataset is applicable to robotics learning and control tasks, particularly robotic arm manipulation and visual perception.